Mathematical Reasoning on AIME 2025 (Pass@32 Metrics)
30.4Avg@32EntroAdv
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| EntroAdvMethod Category=Policy Entropy Method, Backbone=DeepScaleR-1.5B-8k2026.04 | 30.4 | 54.5 | 56.3 | |
| KL-CovMethod Category=Policy Entropy Method, Backbone=DeepScaleR-1.5B-8k2026.04 | 30.1 | 54.9 | 56.5 | |
| EAPOMethod Category=Policy Entropy Method, Backbone=DeepScaleR-1.5B-8k2026.04 | 30.1 | 55.1 | 57 | |
| DAPOMethod Category=Classic RLVR Method, Backbone=DeepScaleR-1.5B-8k2026.04 | 28.6 | 53.6 | 55.1 | |
| ForkingMethod Category=Policy Entropy Method, Backbone=DeepScaleR-1.5B-8k2026.04 | 28.1 | 55.5 | 53.7 | |
| GRPOMethod Category=Classic RLVR Method, Backbone=DeepScaleR-1.5B-8k2026.04 | 27.1 | 49.1 | 53.9 | |
| EntroRegMethod Category=Policy Entropy Method, Backbone=DeepScaleR-1.5B-8k2026.04 | 26.4 | 49.2 | 51.9 | |
| W-REINFMethod Category=Classic RLVR Method, Backbone=DeepScaleR-1.5B-8k2026.04 | 25 | 49.3 | 51.8 | |
| Base-1.5BMethod Category=Classic RLVR Method, Backbone=DeepScaleR-1.5B-8k2026.04 | 24.3 | 51.9 | 50.7 |